Downloadable ContentDownload PDF
Lung cancer detection and classification using SVM
Lung cancer seems to be a common cause of death among people throughout the world. Lung cancer is the leading cancer killer in both men and women in the U.S. In 1987, it surpassed breast cancer to become the leading cause of cancer deaths in women. An estimated 158,080 Americans died from lung cancer in 2016, accounting for approximately 27 percent of all cancer deaths. Early detection of lung cancer can increase the chance of survival among people. The overall 5-year survival rate for lung cancer patients increases from 14 to 49% if the disease is detected in time. Computed Tomography (CT) scans of lungs can be more efficient than X-ray or MRI scans in detecting the presence of cancer. The scanned images of lungs are obtained from LIDC (Lung Image Database Consortium). The scans of twenty patients contain both positive and negative scans I,e. scans with and without tumor. The first step is to segment the tumor affected region from the lungs, for this we use Marker Controlled Watershed Segmentation from the Image Processing Toolbox. The next step is to extract the features using Feature Extraction methods from Computer Vision toolbox of MATLAB. Different extraction methods like GLCM, SURF, MSER and BRISK are used. The features are extracted from cancer detected images only. The data or the features extracted is in the form of matrix. These features are used to train the classifier, Support Vector Machine(SVM). SVM classifier is a supervised machine learning algorithm used as a tool for data classification with advantages in handling data with high dimensionality and a small sample size. The performance of the SVM is observed for each feature as input. Hence, a lung cancer detection system that employs Image Processing Techniques is used to detect the presence of lung cancer in CT- images. In this study, MATLAB is the software used.